Introduction
Accurate assessments of regional- and global-scale changes in the terrestrial biosphere are essential if human impacts on biosphere-atmosphere function are to be understood. There are a myriad of ecosystem attributes to be monitored, but quantifying human impacts necessarily includes an evaluation of vegetation cover and net primary productivity (NPP), as these determine amounts of fuel, fiber, and food for human consumption (Running et al. 1999). A global terrestrial observing system is needed that integrates field-based measurements, flux towers, remote sensing, and ecosystem modeling (Baldocchi et al. 1996, Running et al. 1999, Canadell et al. 1999).
Ecosystem process models that simulate carbon, water, and energy exchange between terrestrial ecosystems and the atmosphere require leaf area index (LAI) and vegetation cover as primary drivers (Landsberg and Gower 1997, Waring and Running 1998), and these must be derived by remote sensing. MODIS (Moderate Resolution Imaging Spectrometer) is the primary high temporal frequency mapping sensor onboard NASA's Earth Observing System (EOS) satellite Terra, launched in December 1999. MODIS is poised to become the most important global mapping sensor ever, as it views the entire Earth's surface every 1-2 days acquiring data in 36 spectral bands at spatial resolutions of 250 to 1000m.
Validation of the global data products derived from MODIS and related sensors is essential to both assess product accuracy and to provide feedback to algorithm developers so the algorithms can be improved. Faced with the challenge of validating global remotely sensed products, NASA formed the EOS Validation Program to assist MODIS (and other) Science and Instrument Teams with product validation. For the Land Science Team (MODLand), research at intensive study sites forms the backbone of the validation plan. These have evolved into what constitutes the MODLand core validation sites network. The sites associated with our current project, BigFoot, are important sites within that network. Each BigFoot site is centered on an eddy flux tower that measures continuous water, energy, and carbon fluxes that can potentially be used to validate MODIS products. However, with their relatively small footprint on the order of
1 km2, nearly equivalent to a single MODIS resolution cell (that in most cases will not perfectly align with the footprint), it is important that the spatial context of flux towers be known.
BigFoot is designed to provide that context using a combination of in situ ecological data, Landsat ETM+ data, and ecosystem models (Cohen and Justice 1999). Moreover, BigFoot maps land cover, LAI, fraction absorbed photosynthetic active radiation (fAPAR), and NPP over a 5 x 5 km area around an eddy flux tower at ETM+ resolution. This means we fully characterizes 25 MODIS cells around a given tower site, and are able to test a number of scaling factors that should reveal possible causes of MODIS mapping errors (thereby providing feedback to algorithm developers). Finally, BigFoot takes important steps to enhance the goals of GTOS (the Global Terrestrial Observing System.
